多学科设计优化分布式计算环境的研究

【作者】 蔡苗；

【导师】 余雄庆；

【作者基本信息】 南京航空航天大学 ， 计算机应用技术， 2003， 硕士

【摘要】 多学科设计优化计算框架是指能实现多学科设计优化方法、包含硬件和软件体系的计算环境。分布式计算是多学科设计优化计算框架的一个重要特征。本文的主要目的就是研究多学科设计优化计算框架中所涉及到的有关分布式计算技术。 本文首先回顾了多学科设计优化计算框架的研究现状，分析和比较了现有的几种分布式计算技术。通过对比分析，我们认为CORBA与JAVA技术是目前实现MDO分布式计算环境较为理想的方法。然后对CORBA的主要概念和工作原理作了阐述。将多学科设计优化方法与CORBA技术相结合，探讨了基于CORBA的多学科优化的分布式系统的构造技术，重点论述了遗留系统的继承，专业领域软件的集成与重用。并以协同优化方法分布式计算环境的建立为例，详细阐述了协同优化方法在分布式计算环境下的实现步骤。通过一个多学科设计优化的算例表明：本文给出的多学科设计优化分布式计算方案是可行和有效的。更多还原

【Abstract】 A framework for multidisciplinary design optimization (MDO) is a problem solving environments that support application of multidisciplinary design optimization. Distributing computing is one of the requirements for such a framework. The purpose of this thesis is to investigate the issues of the distributing computing for the MDO framework.The thesis begins with the review of the requirements for MDO computing frameworks. Several distributed computing technology are presented and reviewed. Comparing different kinds of distributed computing technology, the combination of CORBA with Java is considered to be the best way to realize the distributed MDO environment. Concept on CORBA is presented in somewhat detail. Based on CORBA and the requirements for MDO frameworks, several important issues such as inheriting legacy codes, integrating special software. Collaborative Optimization (CO) is selected as a model for implementation of distributed MDO environment. The detail steps are given to implement distributed environment for Collaborative Optimization. A simple example is used to test the procedure presented in the thesis. The results indicate that the combination of CORBA with Java is able to provide an efficient way implementing the distributed computing environment for MDO.更多还原